As is well known in the art, databases play an important role in modern information technology. For example, as one of significant types of databases, relational database systems are widely used for storing and retrieving various types of information. However, in currently ever-advancing information technology, there are demands for high-concurrency read/write of databases, demands for high-efficiency storage and access of mass data, and demands for high scalability and high usability of databases. In this regard, relational databases have many insurmountable problems. For example, relational databases are not suitable to cope with Web2.0 websites, especially Web2.0 purely dynamic websites of a super-large and high-concurrency SNS type.
In this view, the art has proposed non-relational databases (NoSQL). Non-relational databases are characterized by self-organization, self-management, low cost, high scalability, high concurrency, simple query relation and so on. Specifically, a table in a relational database usually stores a formatted data structure, wherein the composition of each entry field is identical, and even if not every entry needs all fields, the relational database still allocates all fields to each entry. Hence, such a structure is a factor behind the performance bottleneck of relational databases. Unlike a relational database, a non-relational database performs storage in the form of key/value pairs, its structure is not fixed, and each entry may have distinct fields and add some own key/value pairs according to needs. Hence, non-relational databases are not limited to fixed structures, which may reduce some temporal and spatial overheads.
Currently, since non-relational databases have no schema similar to relational databases, it is impossible to perform processing, such as data matching, data integration, data exchange and the like, on relational databases and non-relational databases according to schemas, and further, it is impossible to integrate relational databases and non-relational databases.